[webkit-reviews] review denied: [Bug 18437] Mouse events occur on the wrong element when clip path is applied : [Attachment 20499] First attempt

[bugzilla-daemon at webkit.org]

Eric Seidel <eric at webkit.org> has denied Rob Buis <rwlbuis at gmail.com>'s request
for review:
Bug 18437: Mouse events occur on the wrong element when clip path is applied
https://bugs.webkit.org/show_bug.cgi?id=18437

Attachment 20499: First attempt
https://bugs.webkit.org/attachment.cgi?id=20499&action=edit

------- Additional Comments from Eric Seidel <eric at webkit.org>
Looks OK, but needs more thought than I have brain cells for atm.

First, your doubtful comments about absolute-to-local conversions aren't
needed. :)   Yes in inverse of the absoluteTransform is exactly what will
convert from screen chords to local coords for the element in question. :)

This needs test cases.

Not sure how fast/slow this is.

This doesn't work for TextClips or clips which can't convert to a path object
(which we have a separate bug about fixing somewhere).

I think if you're in the clipping code you should be at least aware of these
other bugs: bug 12571, bug 14015 and bug 18398

why wouldn't "clipContains" just be on RenderObject? (or is there a
RenderSVGObject yet... I guess not).

I guess I'll go ahead and r- this for lack of test case.

The best way to fix this IMO, would be to make sure that any fix moves us
closer to an architecture which we can solve bug 12571 under (Which means
asking the clipper if they contain a point, instead of asking them for their
path first... which I guess you're already sorta doing.)  Maybe I'm just too
tired...